# A/B Testing

{% hint style="info" %}
Available from the Growth plan.
{% endhint %}

### Overview

A/B Testing lets you compare two versions of a demo and understand which one performs better based on **real user behavior**.

Instead of guessing or debating internally, you can run controlled experiments and optimize demos for:

* Higher completion rates
* More CTA clicks
* More leads captured
* Stronger buyer intent

This feature is designed specifically for **marketing and growth teams** using demos as a conversion asset.

{% @storylane/embed subdomain="app" linkValue="xos6sibfdhof" url="<https://app.storylane.io/share/xos6sibfdhof>" %}

***

### Who this feature is for

* Product marketing teams
* Growth & CRO teams
* Demand generation and website teams

If demos are part of your acquisition, activation, or lead capture flow, A/B Testing helps you continuously improve performance.

***

### How A/B testing works (conceptually)

An A/B test compares:

* **Variant A** - your original demo
* **Variant B** - another demo (ideally, a modified, duplicated version of the same demo)

Traffic is automatically split between the two variants, and Storylane tracks how each version performs across key metrics.

Both variants receive **real traffic**, so results reflect actual buyer behavior - not simulations.

***

### Best practice: d**uplicate the demo for variant B**

Before creating an A/B test, we suggest duplicating your original demo, making changes, and using it as variant B.

This ensures:

* Identical demo size and layout
* Clean, comparable results
* Only one variable is being tested

> ⚠️ Changing multiple things at once makes results harder to interpret.

**Recommended:**\
Duplicate → change one thing → test.

***

### What you should test

Marketing teams commonly test:

* Lead form vs no lead form
* Short vs long demo
* CTA at the start vs CTA at the end
* Gated vs ungated experience
* Use-case intro vs feature-first intro
* Video + voiceover vs silent demo
* Strong CTA copy vs soft CTA copy

Always test **one variable at a time**.

***

### Understanding Traffic Split

#### Default: 50/50

Best for:

* Most experiments
* Faster, balanced learning

#### When to Adjust the Split

You may want to change the traffic split when:

* Protecting a high-performing demo
* Testing a major structural change
* Running experiments during high-traffic campaigns

**Example:**

* 80% → proven demo
* 20% → experimental version

This balances learning speed with conversion risk.

***

### Understanding Test Results

#### Clear Winner

One variant consistently outperforms the other across key metrics.

#### No Clear Winner

This is still a valuable outcome.\
It tells you that the tested change did **not** significantly impact performance.

Use these learnings to inform your next experiment.

***

### Ending an A/B Test

When you’re ready, you can end the test to stop traffic splitting.

#### If Variant A Wins

* No action required
* Your existing embedded demo link already points to Variant A

#### If Variant B Wins

* Update the demo link wherever it’s embedded\
  (website, landing pages, campaigns, emails)
* This ensures future traffic goes to the better-performing demo

Ending a test is a deliberate decision and cannot be undone.

***

### Tips for Better Experiments

* Test one variable at a time
* Let the test run long enough to collect meaningful data
* Avoid ending tests too early
* Use results to plan the next iteration

A/B Testing works best as a **continuous optimization loop**, not a one-off task.